NVIDIA’s CEO, Jensen Huang, just dropped a bombshell at Computex 2025. He says the old rule of thumb for tech progress, Moore’s Law, is toast. Instead, he’s pitching the “Huang Law” as the new way to measure innovation.
So, what’s the big deal? Moore’s Law said that the number of transistors on a chip would double every two years. But Huang claims that NVIDIA’s AI chips are moving way faster than that. In fact, their new Blackwell chips are up to 30 times more powerful than the previous generation.
Huang says this is because NVIDIA is working on multiple fronts at once – architecture, chips, systems, libraries, and algorithms all at the same time. This lets them innovate faster and smarter. The “Huang Law” suggests that AI chip performance could double or even triple every year, thanks to improvements in hardware, software, networks, and more.
But some folks are skeptical. They think the “Huang Law” might be more of a marketing gimmick than a real scientific observation. Still, NVIDIA’s progress in AI chips is hard to deny. They’re finding ways to get around the limits of transistor miniaturization, which is what Moore’s Law was all about.
The Numbers Behind the Claim
Let’s take a closer look at what Huang is saying. If the “Huang Law” holds up, it could mean huge leaps forward for AI and computing power. But it’s not just about the tech – it’s also about the business. Can NVIDIA keep up this blistering pace of innovation? And what does it mean for the rest of the industry?
Key Points to Consider
- NVIDIA’s AI chips are advancing faster than Moore’s Law predicted
- The “Huang Law” suggests that AI chip performance could double or triple annually
- This is thanks to improvements in hardware, software, networks, and more
- Skeptics think the “Huang Law” might be a marketing strategy
- NVIDIA’s progress in AI chips is still impressive, regardless of the “law” behind it
The jury’s still out on whether the “Huang Law” is for real. But one thing’s for sure – NVIDIA is pushing the boundaries of what’s possible with AI and computing power. And that’s something to get excited about.